What thermal paste method do you use on your home PC?

[Zodiark1593]

Now that I think on it, I don't even think I properly cleaned off all the old grease from my i5 on my last heat sink mount, just did a quick wipe with the paper towel, threw on the dot of heat grease, and slapped the stock heat sink on.

Winter, temps don't exceed 65c. Quite frankly, I don't think these cpu's care that much how you glop on your paste unless you're overclocking and need every degree or are using crappy paste.

 

[Leyawiin]

Pea on a smooth bottomed cooler and lines along the copper of a direct contact one.

 

[Rakehellion]

Put a glob in the middle and smear it around with the heat sink.

 

[Pontific8]

Always use the rice/peadot technique for the square portion of your cpu it will endeavour to encapsulate the space deftly per square inch.

 

[Puffnstuff]

I use the grain of rice method and let the heatsink spread it form me which has worked well for years. Much more than that and it will ooze over the sides of the cpu lid and make a mess.

 

[pcgeek11]

As evidenced by this thread all methods work...

 

[SteveGrabowski]

Puget Systems has an article about this. The methods all look to have pretty similar results with MX-2, with the worst (thick line using too much) differing from the best (thin X) by only 2C under load on a stock 3770k.

https://www.pugetsystems.com/labs/a...ication-Techniques-170/#ApplicationTechniques

 

[ioni]

An uncooked grain of rice, from my experience, is all you need. Remember that thermal paste is meant to fill in the space caused by microscopic valleys and ridges on the mating metal surfaces. You want it to be able to spread nearly to the edge of the IHS in an extremely thin film. If it's squishing out over the edges, you've used too much.

People always say this, but I could never see how such a small amount of paste could fill the whole square. You'd need so much pressure. I use a line, which is probably the volume of 2 peas, and when I took my HS off to change the CPU, I had a very thin layer of grease across most of the square. That much didn't even fill the entire square!

 

[Yuriman]

People always say this, but I could never see how such a small amount of paste could fill the whole square. You'd need so much pressure. I use a line, which is probably the volume of 2 peas, and when I took my HS off to change the CPU, I had a very thin layer of grease across most of the square. That much didn't even fill the entire square!

Use whatever it takes to spread across the IHS without going over the edges. I've always lapped my contact surfaces, so it doesn't take much, and Intel's IHS's are smaller than AMD's. The die is only a tiny fraction of the IHS though, and you only really need coverage of the IHS over where the die is:

 

[Phynaz]

I'm not sure about that. The IHS is meant to spread the generated heat over a larger area, therefore your heatsink should have good contact over that entire area to remove that heat.

 

[Arachnotronic]

I do the pea dot.

 

[Yuriman]

I'm not sure about that. The IHS is meant to spread the generated heat over a larger area, therefore your heatsink should have good contact over that entire area to remove that heat.

No, it's meant to protect the die. A metal IHS does no better job at transferring heat laterally than the metal base of a heatsink. Worse, in fact, which is why many enthusiasts remove them.

 

[Teabaghead]

I usually use a shovel to apply the paste

 

[MrTeal]

No, it's meant to protect the die. A metal IHS does no better job at transferring heat laterally than the metal base of a heatsink. Worse, in fact, which is why many enthusiasts remove them.

The larger surface area does help to reduce the heat flux density though. For HEDT, there is benefit from going from the small die through a high conductivity interface (solder) to the heat spreader, and then having a larger area interface with the poorer TIM to the heatsink.

On of the things I keep meaning to do when I have more time on my hands (Ha!) is to heat the top off an FX CPU and then solder the die directly to a waterblock, and log the core to water delta before and after.

 

[lord_emperor]

Puget Systems has an article about this. The methods all look to have pretty similar results with MX-2, with the worst (thick line using too much) differing from the best (thin X) by only 2C under load on a stock 3770k.

https://www.pugetsystems.com/labs/a...ication-Techniques-170/#ApplicationTechniques

I was going to link this exact article and add the caveat that while the test conditions look like they were setup very carefully, with only one test of each application method the variance from human error could easily account for the differences.

I used a "grain of rice" of Arctic Silver 3 (just finished off my tube...) last time I put my heatsink on and I'm not going to remove it just to change to an "X".

 

[Puffnstuff]

I usually use a shovel to apply the paste

Your sarcasm sir is in great form. :biggrin:

 

[bononos]

I was going to link this exact article and add the caveat that while the test conditions look like they were setup very carefully, with only one test of each application method the variance from human error could easily account for the differences.
.......

Thats true but considering nearly every method in a dozen tests either had large and small bubbles, its fair to say that simply slapping on paste isn't right for proper application.

 

[bystander36]

What I took away from that test was there was only 2C difference between the best and worst application.

 

[bradly1101]

I just use the credit card spread, and never had temp. problems.

 

[pcgeek11]

What I took away from that test was there was only 2C difference between the best and worst application.

This.

 

[MongGrel]

I just use the credit card spread, and never had temp. problems.

That is even just old.

I used to spread it carefully with a razor blade.

Put on a Pea sized dot on a solid bottom, just attach it with the springs and it will spread out on its own without air pockets.

 

[swampi800]

Pea size definitely or cross. IHS is a heat spreader besides protecting the die as the thermal solution I think that contact is always the primary issue – more/better contact more performance too little paste and it will impact your performance in my experience. I think it depends if your heat sink and IHS are generally flat not so much, overly warped sink or IHS then use more.

OMG
Google edstechreport How Long Does it Really Take To Break-In The Arctic Silver 5 Thermal Paste






1.5 million spam posts wow

 

[Puffnstuff]

The whole purpose of the TIM in the first place is to fill the gaps between the spreader and the bottom of the HSF to aid in heat transfer and nothing more. When you cake it on you're actually reducing the effectiveness of this transfer. The grain of rice method has worked fine for me since I started building my own systems and when I remove the hsf I find that the base was adequately covered which is all it is supposed to do.

 

[bononos]

What I took away from that test was there was only 2C difference between the best and worst application.

Which was measured probably on the same day as the thermal paste application itself.

The temps could worse in the medium and long term with the pump out effect on an already mildly faulty application on the first day compared to a cpu with a proper application of paste.

 

[swampi800]

The roll call discussion of the day concerning technique i guess is one of sensibility rather than meticulous pedanticism.